HiGHS

Provides an MCP-enabled HiGHS solver for LP/MIP optimization with flexible problem specification and robust outputs.

typescript

12

GitHub Stars

typescript

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "wspringer-highs-mcp": {
      "command": "npx",
      "args": [
        "highs-mcp"
      ]
    }
  }
}

You can run the HiGHS MCP Server to solve linear and mixed-integer optimization problems through a standard MCP interface. It provides fast optimization using the HiGHS solver and supports LP, MIP, QP (convex), and multi-objective problems, with flexible problem definitions and robust output.

How to use

Start the MCP server locally and connect your MCP client to solve optimization problems. You can run it directly and then send problems in the MCP format through your client. The server exposes a single tool that accepts problem definitions, runs the HiGHS solver, and returns the solution and supporting values.

How to install

Prerequisites: ensure you have Node.js version 16.0.0 or newer installed on your machine.

Install the HiGHS MCP Server package from npm.

npm install highs-mcp

If you prefer to build from source, clone the project, install dependencies, and build.

git clone https://github.com/wspringer/highs-mcp.git
cd highs-mcp
npm install
npm run build

Run the server directly using the runtime command.

npx highs-mcp

Claude integration configuration

To enable Claude to use the HiGHS MCP Server, add the server configuration to Claude's MCP settings.

{
  "mcpServers": {
    "highs": {
      "command": "npx",
      "args": ["highs-mcp"]
    }
  }
}

Configuration and usage notes

The server accepts optimization problems through a standardized tool API. You can specify objective sense, linear and quadratic components, variables with bounds and types, and linear constraints in either dense or sparse formats. The solver options let you tune time limits, tolerances, and algorithm choice.

Troubleshooting and tips

Verify your Node.js version is compatible (Node.js >= 16). If the server fails to start, ensure dependencies are installed and the build step completed if you built from source. If you encounter performance issues, adjust solver options such as time_limit, threads, and simplex or ipm settings to suit your problem size.

Available tools

optimize-mip-lp-tool

Executes LP/MIP optimization through the MCP interface, returning primal solution, dual values, and reduced costs.